© Copyright Australian Museum, 2001 Records of the Australian Museum, Supplement 27 (2001): 67–74. ISBN 0 7347 2305 9 Shell and Bone Artefacts from the Emily Bay Settlement Site, Norfolk Island LYN SCHMIDT1, ATHOLL ANDERSON1 AND RICHARD FULLAGAR2 1 Department of Archaeology & Natural History, Research School of Pacific and Asian Studies, Australian National University, Canberra ACT 0200, Australia [email protected] [email protected] 2 School of Geosciences, University of Wollongong, Wollongong NSW 2500, Australia [email protected] ABSTRACT. Amongst molluscan shell from the Emily Bay site were 40 specimens of fragmented bivalves, especially of Gari livida, which were examined for evidence of their use as artefacts. Experiments using modern specimens of the same taxa showed that it was impossible to define deliberate breakage sufficiently clearly to define shell tools on that criterion. Analysis of usewear by microscopic inspection of edges was the main discriminant adopted. In addition vegetable residues were identified on several edges. These means identified 19 pieces as tools, which had been used mainly for scraping soft materials. Two other tools were identified by morphology. A small assemblage of bone and marine ivory artefacts was also recovered from Emily Bay. Most were pieces of fishing gear. SCHMIDT, LYN, ATHOLL ANDERSON AND RICHARD FULLAGAR, 2001. Shell and bone artefacts from the Emily Bay settlement site, Norfolk Island. In The Prehistoric Archaeology of Norfolk Island, Southwest Pacific, ed. Atholl Anderson and Peter White, pp. 67–74. Records of the Australian Museum, Supplement 27. Sydney: Australian Museum. Shell artefacts direct relationship between usewear and residues has been established for any of these artefacts. This paper attempts The shell artefacts referred to here are not of the formal to address this issue by an examination of usewear and kinds found throughout Oceania (Poulsen, 1970), but rather residues on the worked edges. No signs of deliberate informal, flaked shell pieces. Fragments of worked shell modification or use were observed on any gastropod shell appear in Pacific sites from early Lapita in the west (Spriggs, recovered from the site at Emily Bay. In contrast, all of the 1991; Kirch, 1987) to late sites in the east (Kirch, 1989), bivalve shell (40 pieces) was highly fragmented, often in thus possessing a very wide geographic and temporal ways which suggested deliberate human modification. In distribution. However, they have been considered to yield order to cast some light on the processes which might have no information about cultural sequences and so little time been involved in shell modification at the site, experimental has been devoted to their analysis. They have been described breakage was conducted with material from the same variously as worked shell, shell fragments, shell scrapers molluscan species, and the archaeological specimens were or shell knife fragments. Smith (1999: 284) notes that no examined microscopically for edge wear and residues. 68 Records of the Australian Museum, Supplement 27 (2001) Shell breakage experiments. Breakage pattern experiments cutting edges needing no further modification, and that flake were carried out on modern specimens of the main bivalves removals were unpredictable according to planes of in the Emily Bay assemblage, Gari livida, Pinctada weakness within the shell matrix. Both these findings were maculata and Mactra rufescens. The experiments aimed to confirmed in our study. simulate the two main processes other than deliberate Application of weight by crushing or trampling resulted modification likely to have affected the shells. These are either in shells with the valve intact but flakes removed impact breaks and trampling. Impact breaks may have been along the margins, or a clean, lateral break separating the caused either by meat extraction or shell discard processes, inferior and superior portions of the valve. These distinctive or by subsequent natural forces, such as storm damage. patterns are also illustrated in Fig. 1. Secondly, there is the post-depositional effect of trampling While these experiments produced all of the shell over the site. Twenty bivalves were included in the breakage modification features apparent in the archaeological experiments (dropping shells onto a hard floor) and in assemblage, it was not possible to distinguish clearly experimental trampling. An additional five shells were between the agencies in any particular case. Sharp working snapped between thumb and forefinger to replicate edges result from all agencies and it is not possible to deliberate breakage to form a clean working edge. demonstrate that deliberate breakage aimed at producing a These experiments produced breakage patterns similar working edge was responsible in most cases. Abrasion along to those in the Emily Bay assemblage. Deliberate snapping edges is also not diagnostic as it might occur through the between thumb and forefinger resulted in fracturing that action of wind or water and sand. The only valves that can could not be differentiated from that caused by discard definitely be classed as tools are those that contain retouch throws (impact) or smashing to extract meat. Both snapping and/or usewear other than slight abrasion along their and impact forms of modification resulted in a segment margins. These experimental results conform with those retaining some portion of the original valve edge with a found in earlier work (Spennemann, 1993: 80) on Anadara sharp internal edge (Fig. 1). Edges were very angular and shell when the different strength of the shell structure for sharp in G. livida and M. rufescens due to the tendency in the species involved is taken into consideration. these species for the material to split along cleavage planes in the shell structure. Without any further modification a Usewear analysis. Usewear on shell artefacts has been very sharp working edge is thereby obtained. Previous shell studied on a number of occasions (Attenbrow et al., 1998; flaking experiments (Cleghorn, 1977) had demonstrated that Barton and White, 1993; Cleghorn, 1977; Cooper, 1988; initial breakage of shell produced pieces that exhibited sharp Fullagar, 1986; Fullagar et al., 1992; Kamminga, 1982; Figure 1. Shell breakage patterns. Examples (left) show typical damage from trampling in which the hinge remains intact and flakes are removed around much of the periphery. Examples (right) show typical snap patterns in which there are long, clean breaks and a portion of the valve edge remaining. Schmidt et al.: Shell and bone artefacts 69 Spennemann, 1993; Toth and Woods, 1989). However, there the presence of usewear or retouch were required to class a has been relatively little systematic experimentation with specimen as a shell tool. Three artefacts had been formed shell and the interpretation of wear patterns can only be by quite deliberate flaking and grinding irrespective of tentative until a more extensive range of studies is additional evidence of use. These atypical artefacts are undertaken. In addition it is rare in archaeological reports pictured in Fig. 2. for small, informal shell tool types to be recorded (Janetski, Any artefacts that exhibited usewear traces were then 1976; Lima et al., 1986; Reiger, 1981; Schrire, 1982). In examined at magnifications up to ×500, using an Olympus reports where these tools are pictured they appear similar metallographic microscope with vertical incident light. to the modified shell pieces recovered from Norfolk Island Sketches and observations of usewear were recorded for (e.g., Lima et al., 1986). each shell. A combination of diagnostic use traces is A search for both macroscopic and microscopic usewear necessary to distinguish usewear because of the impact of was carried out on the shell artefacts from the Emily Bay weathering on shell, which is more vulnerable than stone. site. Wear was defined as consisting of edge fracturing, Interpretations of shell artefacts as utilized tools have been striations and edge dulling or rounding, as defined by classified into four levels of confidence dependent on the Kamminga (1982). These criteria were used to establish incidence of diagnostic traces and their combination with use, along with deliberate modification by flaking of an unstable thin edges which are prone to incidental damage: edge. Barton and White (1993) found that fractures on a 0 possible use: shaping but no wear traces shell matrix did not possess the clear characteristics of feather, snap and step fractures found on stone. A further complication 1 possible use: rounding and scarring in combination results from the findings that edge rounding and striations have with weathered or unstable edges also been found to be the result of wave action (Spennemann, 2 probable use: rounding and scarring 1993; Claasen, 1998; Toth and Woods, 1989). 3 definite use: clear distinctive usewear Analysis of the archaeological assemblage. All of the 40 Following the usewear analysis, edges that showed use pieces which appeared to have been artefacts were examined traces were examined for residue analysis. Survival of macroscopically and in magnifications up to about ×50 residues is uncommon in the archaeological assemblage. using a Zeiss stereomicroscope. Since it is not possible to Films and fragments of unidentified plant tissue are present distinguish deliberate from natural processes in breakage, and starch grains were identified on two of the shell edges. Figure 2. Atypical shell artefacts. 70 Records of the Australian Museum, Supplement 27 (2001) Table 1. Shell artefacts from Emily Bay settlement site. Trench/Square/ species use type of usewear and modification weight length Spit class (g) (mm) EB96:10 A2/1 G. livida 1 All margins removed to get semi crescentic 0.1 15.2 shape. EB96:10 A2/2 M. rufescens 2 Slight rounding and denticulate edge on 7.3 0.9 19.8 mm length. EB96:10 A2/2 M. rufescens 3 Section of usewear along lateral shell edge 2.2 31.6 13.4 mm length with scarring and rounding of edge. EB96:10 A2/4 ? 3 Section of usewear along lateral shell edge 2.7 30.4 24.5 mm length with striations and rounding of margins.